I'd think first there would need to be listening tests to find out if those "dynamic" speakers sound that way regardless of listening (and listener) conditions and recordings. From there it should be easier to proceed.
Some days I do experience "dynamics" differently although nothing obvious was changed.
Last edited:
Assuming this being a real system, how would you estimate
consistency of dispersion, say above 600 Hz ?
Upper set of curves around 00 degrees off axis
Middle set of curves around 45 degrees of axis
Lower set of curves around 65 degrees of axis
consistency of dispersion, say above 600 Hz ?
Upper set of curves around 00 degrees off axis
Middle set of curves around 45 degrees of axis
Lower set of curves around 65 degrees of axis
Attachments
Last edited:
Hi Earl, ra7, all
On the MTF comment I made, the message was supposed to be something like this. MTF’s are used to calculate the STIpa measurement which predicts intelligibility.
The MTF itself is a measurement of the dynamic information reaching the listener and is affected by many aspects of loudspeaker performance.
All other “loudspeaker” things being equal, one thing that harms / reduces the dynamic of what reaches the listener is how much reflected sound there is relative to the direct sound. This would not apply to “drums in a room” because THAT is the original signal but rather would apply to drums played through speakers in a room, or as it was intended, intelligibility or random words in larger spaces.
It has been my observation that the preservation of the recorded stereo image also seems to depend on how much of the dynamic is preserved at the listening position and that IS (all other things being equal) related to directivity and position when in a room. . Understand I am not talking about enjoy ability, we can enjoy a huge range of radiation styles, I am talking about part of what governs the ability to paint a picture between the two speakers where (if the recording has it) can make sources appear in front of you where there are no loudspeakers.
If one wants to strip away the room related parts and listen to the loudspeaker part alone, set up outside or sit very close to the speakers near the center of a room.
I think one can also draw a little bit from the old days
Hopkins stryker” equation is an early attempt which tried to correlate a number of factors to predict intelligibility. Intelligibility goes down as you increase the number of sources, as you decrease the Q, if they are not aimed at the listener and so on.
In the mid 80’s I presented an AES paper on the elimination of power compression that was done by using a forced air cooling system that was powered by a small bit of the incoming amplifier power. That was as applied to the Servomotor driven subwoofers we sold at Intersonics at the time. Later a friend used the same approach under license on conventional loudspeakers with good results. Steve is now head of R&D at 18sound but his aes paper is here;
AES E-Library Power Compression Elimination in Moving-Coil Loudspeakers by the Use of Fan-Actuated Forced-Air Cooling
On “power compression” it’s been a long time since I was working on this but there is a part of it which may be important. If memory serves, at about 230C the resistance of copper has doubled which would obviously be a “change”. Some of today’s high power voice coils can get considerably warmer than that and not fail. That is in contrast to the epoxy adhesives in the 70’s when loudspeakers didn’t have power compression….because the wire would come unglued at about 150C.
At the same session as my paper, Doug Button (JBL at the time) presented a paper looking even higher temperatures and found that when the VC temp is high enough, a further increase in input power resulted in a reduction in output.
Feeding a loudspeaker pink noise and using a time vs SPL recoding, one sees the effect of PC by it’s reduction in SPL over time. AS pointed out, there is a thermal times constant, actually several that govern the short and long term heating and cooling times. The voice coil is suspended in air however, the same thing our houses are insulated with and being a very poor conductor of heat is why drawing air across the conductor has a large effect as Steve found in the old days on his conventional drivers.
While we can hear changes in level, what may well be much more audible is how the change in Rdc with heat, changes the frequency response of a loudspeaker. Where ever the impedance is lowest, that is where one sees the greatest reduction in SPL and where highest, the least change.
That frequency response modulation may well be more audible than the reduction of average spl and depends on the driver, how it’s used and so on.
Earl, I would love to hear your system and I may get to go back to Michigan state this season (some talk of seeing a game) so if I am, I will get a hold of you beforehand.
I don’t have these where they can be downloaded yet (trying to get them on our website somewhere) but If you have headphones on your computer or can play your computer through your system I would be curious about your thoughts on a recording I made n the 4th of July with a “capture” system I have been playing with.
With speakers like we are talking about, this comes across pretty well in a living room and somewhat different but good on headphones. Stay cool!
Best,
Tom
Try the parade here;
https://soundcloud.com/tomdanley
On the MTF comment I made, the message was supposed to be something like this. MTF’s are used to calculate the STIpa measurement which predicts intelligibility.
The MTF itself is a measurement of the dynamic information reaching the listener and is affected by many aspects of loudspeaker performance.
All other “loudspeaker” things being equal, one thing that harms / reduces the dynamic of what reaches the listener is how much reflected sound there is relative to the direct sound. This would not apply to “drums in a room” because THAT is the original signal but rather would apply to drums played through speakers in a room, or as it was intended, intelligibility or random words in larger spaces.
It has been my observation that the preservation of the recorded stereo image also seems to depend on how much of the dynamic is preserved at the listening position and that IS (all other things being equal) related to directivity and position when in a room. . Understand I am not talking about enjoy ability, we can enjoy a huge range of radiation styles, I am talking about part of what governs the ability to paint a picture between the two speakers where (if the recording has it) can make sources appear in front of you where there are no loudspeakers.
If one wants to strip away the room related parts and listen to the loudspeaker part alone, set up outside or sit very close to the speakers near the center of a room.
I think one can also draw a little bit from the old days
Hopkins stryker” equation is an early attempt which tried to correlate a number of factors to predict intelligibility. Intelligibility goes down as you increase the number of sources, as you decrease the Q, if they are not aimed at the listener and so on.
In the mid 80’s I presented an AES paper on the elimination of power compression that was done by using a forced air cooling system that was powered by a small bit of the incoming amplifier power. That was as applied to the Servomotor driven subwoofers we sold at Intersonics at the time. Later a friend used the same approach under license on conventional loudspeakers with good results. Steve is now head of R&D at 18sound but his aes paper is here;
AES E-Library Power Compression Elimination in Moving-Coil Loudspeakers by the Use of Fan-Actuated Forced-Air Cooling
On “power compression” it’s been a long time since I was working on this but there is a part of it which may be important. If memory serves, at about 230C the resistance of copper has doubled which would obviously be a “change”. Some of today’s high power voice coils can get considerably warmer than that and not fail. That is in contrast to the epoxy adhesives in the 70’s when loudspeakers didn’t have power compression….because the wire would come unglued at about 150C.
At the same session as my paper, Doug Button (JBL at the time) presented a paper looking even higher temperatures and found that when the VC temp is high enough, a further increase in input power resulted in a reduction in output.
Feeding a loudspeaker pink noise and using a time vs SPL recoding, one sees the effect of PC by it’s reduction in SPL over time. AS pointed out, there is a thermal times constant, actually several that govern the short and long term heating and cooling times. The voice coil is suspended in air however, the same thing our houses are insulated with and being a very poor conductor of heat is why drawing air across the conductor has a large effect as Steve found in the old days on his conventional drivers.
While we can hear changes in level, what may well be much more audible is how the change in Rdc with heat, changes the frequency response of a loudspeaker. Where ever the impedance is lowest, that is where one sees the greatest reduction in SPL and where highest, the least change.
That frequency response modulation may well be more audible than the reduction of average spl and depends on the driver, how it’s used and so on.
Earl, I would love to hear your system and I may get to go back to Michigan state this season (some talk of seeing a game) so if I am, I will get a hold of you beforehand.
I don’t have these where they can be downloaded yet (trying to get them on our website somewhere) but If you have headphones on your computer or can play your computer through your system I would be curious about your thoughts on a recording I made n the 4th of July with a “capture” system I have been playing with.
With speakers like we are talking about, this comes across pretty well in a living room and somewhat different but good on headphones. Stay cool!
Best,
Tom
Try the parade here;
https://soundcloud.com/tomdanley
Yes, it sure would be nice to do all sorts of cool subjective tests. I hope someone can do that someday.
From an industry standpoint it does not seem like doing these kinds of tests nets anything positive. Academically there isn't any financial support for it. And for individuals it is just plane impossible to do anything that wouldn't be picked apart for its lack of controls.
The power / thermal control discussion is interesting, albeit a little bit off topic. Definitely a worthwhile discussion though. It is hard to know if "micro-compression" is what makes high-efficiency speakers sound so much better at low power levels. I kind of doubt it, since the typical high-efficiency DI-matched two-way speaker uses a 600 watt midwoofer and 100 watt compression driver, and they are pretty loud at 10 watts and 1 watt respectively. I think probably something else is going on, maybe just the fact they have greater dynamic range.
At the full-tilt end of the scale though, I do think there is a sort of compression modulation effect. Surely, we can measure compression easily, just a difference of the SPL measured and the expected value extrapolated from lower power levels. For example, if the SPL at 1W/1M is 100dB, and we drive the speaker with 1000 watts, we might expect 130dB. When we measure 125dB instead, we see there is 5dB compression. Beyond this - and to the point of compression modulation - When I measured woofers near their thermal limits, they tended to steadily decrease SPL with a constant drive level after a sine wave burst started. This can be called a compression modulation, although the speed of the modulation for woofers is fairly slow, on the order of several seconds, because the thermal mass is large. Smaller drivers would undoubtedly enter compression faster.
I also wanted to say that the whole concept of thermal dissipation for loudspeakers is somewhat misunderstood, in my opinion. The amount of heat generated by eddy currents in the magnet is significant, and largely overlooked. This raises the local ambient temperature of the voice coil, which is undesireable. We can always use forced air convecton to cool something, such as is the case with air cooled Volkswagon motors. The faster the airflow, the greater the cooling. But it isn't the only way to cool the speaker, and in fact, isn't even the best way. It's just the one that intuitively seems beneficial, and it also has the benefit of simplicity, which is a virtue unto itself. So while I wouldn't discount this method, I would suggest that other methods should also be used, where maximum thermal control is required.
Back to the "uniform-directivity" topic. I noticed some discussion on a thread about horn profiles and throat adapters that is relevant here. This other thread also wrestles with the potential problem of "throwing the baby out with the bathwater" when designing waveguide/horns. Check it out:
At the full-tilt end of the scale though, I do think there is a sort of compression modulation effect. Surely, we can measure compression easily, just a difference of the SPL measured and the expected value extrapolated from lower power levels. For example, if the SPL at 1W/1M is 100dB, and we drive the speaker with 1000 watts, we might expect 130dB. When we measure 125dB instead, we see there is 5dB compression. Beyond this - and to the point of compression modulation - When I measured woofers near their thermal limits, they tended to steadily decrease SPL with a constant drive level after a sine wave burst started. This can be called a compression modulation, although the speed of the modulation for woofers is fairly slow, on the order of several seconds, because the thermal mass is large. Smaller drivers would undoubtedly enter compression faster.
I also wanted to say that the whole concept of thermal dissipation for loudspeakers is somewhat misunderstood, in my opinion. The amount of heat generated by eddy currents in the magnet is significant, and largely overlooked. This raises the local ambient temperature of the voice coil, which is undesireable. We can always use forced air convecton to cool something, such as is the case with air cooled Volkswagon motors. The faster the airflow, the greater the cooling. But it isn't the only way to cool the speaker, and in fact, isn't even the best way. It's just the one that intuitively seems beneficial, and it also has the benefit of simplicity, which is a virtue unto itself. So while I wouldn't discount this method, I would suggest that other methods should also be used, where maximum thermal control is required.
Back to the "uniform-directivity" topic. I noticed some discussion on a thread about horn profiles and throat adapters that is relevant here. This other thread also wrestles with the potential problem of "throwing the baby out with the bathwater" when designing waveguide/horns. Check it out:
Here's my stab at "uniform directivity" consisting of a SEOS-24, BMS coaxial CD, and an AE TD-15M in a resistive enclosure. I only have a rudimentary passive crossover in place, so it's not near perfect yet.
I would like to keep a similar pattern to 20hz, but that's pretty difficult in a living room setting. I have a little more experimenting to do, but if things don't work out, I have a pair of AE TD-15S I can use from 20-200hz.
I would like to keep a similar pattern to 20hz, but that's pretty difficult in a living room setting. I have a little more experimenting to do, but if things don't work out, I have a pair of AE TD-15S I can use from 20-200hz.
That's kind of hard to read. Maybe post it next to a regular graph of the on-axis response, or normalize it.
Looks easy to read to me. The transition from yellow to green is the -6dB point, right?
It seems pretty tight, if I read it correctly.
It seems pretty tight, if I read it correctly.
Sure, for example, the -6dB point narrows significantly from 3-4k. The on-axis response is also dropping. Is the pattern pinching, or is it purely the on-axis response doing that? Lessee... the on-axis goes from dark red to light orange, while the -20° goes from dark orange to slightly yellowish green... How is this easy? 😛
It's easy to judge the entire system, but I figured the idea was to check out the directivity while keeping in mind the on-axis response isn't finished?
The dip on axis is due to my quick crossover, I put something together to protect the drivers.
Last edited:
I'm not confused, just complaining that the rough response makes it too hard to see directivity, so a normalized graph or just a boring old bunch of lines on a response plot would be far easier. But, that's the same graph.
I'll see what I can pull up on my measurement laptop or will re-measure later this week and will post a normalized graph.
Cool. I am actually interested in your results, not entirely just being a ****. Where are you crossing the woofer?
That shouldn't affect the dispersion, which goes from around 40 degrees at 1000 Hz, to 60 at 3000, and back to only 30 degrees at 4500 Hz, large, non-uniform changes in the two octaves where hearing is generally most sensitive.
Axial normalization highlights directivity, but creates a completely false sense of the overall response. And if the listening axis is not the central axis then normalizing to this axis is completely wrong.
I have no trouble reading the plot (you should get used to it because it is the future I am sure), but my biggest complaint with the plot is its obvious heavy smoothing as no real system would look like that. See my website if you want to see what high resolution data looks like. Nothing like what was shown. The details do matter - a lot.
I have no trouble reading the plot (you should get used to it because it is the future I am sure), but my biggest complaint with the plot is its obvious heavy smoothing as no real system would look like that. See my website if you want to see what high resolution data looks like. Nothing like what was shown. The details do matter - a lot.
- Status
- Not open for further replies.